OPTIMIZATION OF SOME NOISY NONLINEAR SYSTEMS
Abstract
The presence of noisy nonlinearities in the transmission of signals in a system whose function is either signal recovery (estimation) or signal detection is considered. The system is composed of a nonlinear operator followed inseparably by an additive stochastic noise source and interspersed between two linear timeinvariant filters. For the estimation function, two types of signal are considered: a sinusoidal signal of random phase; and a more general stochastic signal of second-order stationarity. For the sinusoidal case, it is shown that infinite prefilter gain is optimum for v-th-law and error function type devices, while for the stochastic case, specification of the optimum postfilter is derived with the same result for prefilter gain. For the detection function, a narrowband signal of known form is assumed. The postfilter gain is determined by matched-filter theory, and it is shown that prefilter gain should be infinite, though the conditions are relatively restrictive for the results. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1961
- Accession Number
- AD0265554
Entities
People
- Robert O. Harger
Organizations
- University of Michigan